Banks Draw Lines Around Customers’ AI Agents

Banks are drawing clearer boundaries around the use of artificial intelligence agents by customers, according to a PYMNTS.com report. The issue centers on how financial institutions handle AI tools that can act on a user’s behalf in banking and payment-related interactions.

AI agents are software systems designed to carry out tasks with varying degrees of autonomy. In financial services, that can raise questions around authentication, customer consent, data access, transaction controls and responsibility when an automated system initiates an action or communicates with a bank.

For banks, the growing use of AI agents creates a need to distinguish between information requests and actions that could affect accounts, payments or sensitive customer data. Existing banking safeguards typically rely on verified identities, permissions and monitoring controls, making autonomous or semi-autonomous software a new operational consideration.

The debate reflects a broader shift as generative AI moves beyond chat-based assistance toward tools intended to complete multi-step tasks. Financial institutions are likely to focus on maintaining security and customer protections while determining how, and under what conditions, customer-controlled AI agents can interact with their systems.